What works? First, we find that this kind of medium gives students and instructors what both feel is needed where grammar instruction is concerned: a quick fix. But beyond the backgrounding rather than foregrounding of grammar, it must be said that acquaintance with the software can be shown to be an effective means of getting students to learn the terminology of the discipline, something that makes their understanding of it, their ability to communicate about language, and their ability to teach others much, much stronger.
Students shared the software readily too, either working in pairs in labs or later on when recommending it to others. The immediate feedback was the main selling feature, but I think the level of recognition needed to spot the correct answer was lowered here from its level on the printed page, something that mirrored the kind of knowledge that grammar is. It is a skill of editing, not of composing, and the emphasis on recognition, rather than reconstruction, makes that clear. The fact that students worked at their own pace too helped to background the grammar instruction. And it meant that the level of social embarrassment sometimes felt in the teaching of grammar did not come into play here.
My experience in coming up with the material needs comment here. From the software developer’s end of things, the project is driven by the exploration of interactive technology. From the content developer's point of view, though, the medium can be a hard mistress. I found the demands of the software more exacting and more rigid than those of the printed page. As a result, the most successful exercises in our cooperative venture were those that illustrated the few hard and fast points about writing that needed to be made. Places where usage made the area not so cut and dried were more difficult. The written exercises I had produced in composition textbooks succeeded by virtue of their flip tone, their sense of humor, and their approach to correctness. This medium seemed to demand less casual responses and a more conservative approach. Hence, computer software, as I learned to create it using MILE, was best suited to fine points of grammar that were easily decided and much more difficult to adapt to stylistic issues of composing. It could, in other words, make better editors, but perhaps not better composers, of its users. And this not just because the software's best form was the multiple choice question, but perhaps also because we demand that computers provide us with the right answer--in absolute terms, in a way that transcends the vagaries of human style. --Fine, then. That at least means that more time in the classroom can be devoted to the composing process, and students will still be able to make good use of this medium to teach them things about language that we need them to know, such as how to identify a main clause and a subordinate clause.
I had often suspected that insistence on grammar was strongest in those who liked to fall back on something rigid. Here the politics of language instruction became more obvious to me than they had ever been before. How well, in fact, is the material suited to the medium? That is the question that we felt had to be asked before the medium was adopted purely for its efficiency at doing some things. Our questionnaires and gathered student results indicate that, in the context of our first-year Multimedia students, both performance and attitude improved as a result of working with the software.